MOTM

I'm not sure if it's okay to post gee-whiz module wishes here, but so
far it seems to be, so I'm going to.  Be aware that I have no real
understanding of what I'm talking about.  If "freshman mistakes" really
annoy you please accept my sincere apologies and just ignore this
message.

Here is one of the modules I'd like but which does not yet exist.
(Actually I want four of them). I don't even know if such a thing can
exist.

The module would be MOTM format, 1U wide, with 4 jacks, three knobs, and
one three-position switch. The jacks would be TRIG IN, TRIG OUT, CV OUT,
and GATE OUT. 

TRIG IN would connect to a usual piezo drum trigger, and ideally would
also work with regular audio in (as is the case with the equivalent to
TRIG IN on many analog drum synths), as well as a regular clock in (such
as a square LFO).

TRIG OUT would generate a trigger pulse suitable for a MOTM-800 EG,
MOTM-100 NOISE/S&H, or other input that makes we Pd programmers think of
a bang. CV OUT would generate a CV based on how hard you squeezed
("squoze?") the piezo (or how loud the audio hit was, assuming that can
be made to happen). GATE OUT would generate a GATE suitable for a
MOTM-800 EG.

The switch would select whether how hard you squeezed controlled GATE
length, CV, and perhaps it might have a middle position in which both
GATE length and CV were so controlled.

The three knobs are SENSITIVITY, GATE TIME, and CURVE. SENSITIVITY
adjusts how much signal you get from the TRIG IN jack. This is a crucial
feature for drum trigger kits, since each drum has a different
sensitivity and it's good to tune them so they're all similar. GATE TIME
is obvious. CURVE controls a response curve that translates how hard you
hit the piezo into how big of a CV and/or GATE you get out. Some like
this very linear, some like it exponential. Exponential CV response
would be especially useful for driving the pitch CVs of VCOs. I have no
idea if this "CURVE" feature is possible.

Obligatory "because this is MOTM" style over-the-top feature, if you're
into that: each of these has *rear panel* connectors and trim pots for
three others. These are used to utterly cancel crosstalk when the piezo
triggers are mounted to the same physical object. Trust me, drummers
love to calibrate against crosstalk for hours. The more difficult it is
to reach these pots, the better. ;)

Why? Well, here's the reason why this module is so desirable. Right now,
if you want to play synthesized drums, you have two choices. One is
MIDI. The shortest MIDI message is somewhere around 1ms of transmit
time, which is only barely noticeable to a good fast drummer. It's
really hard to keep latency (time between physical contact with the
sensor and sound production) down this low, however. I can produce MIDI
in this much time from a Zendrum or an old Alesis D4, however MIDI then
has to get processed by some sound engine to produce sound, and usually
that phase is not only noticeably long (> 5ms), but non-deterministic.
To me, it feels a lot nicer to play into my sampler at a constant 5ms
latency than it is to play into my FM synth with non-deterministic
latencies from 1-12ms. To a fast accurate drummer unlike myself, either
is unacceptable.

The other choice is to get an SDS-8, drumfire, or some other analog drum
synth. Latency basically goes away, which is wonderful. But, analog drum
synths (at least that I've ever seen) are all heavily normalized.
They're also all out of production.

Applications: connect TRIG and GATE to MOTM-800 for control of synth
voice. Connect TRIG to MOTM-101, Arrick sequencer, MILTON, or other
clocked devices for what I call, to coin a phrase, "rich man's tap
tempo". With GATE width control, could also give "pulse width
modulation" control to the SQUARE output of an LFO.  If you ever find
something that pumps out clocks, you could use this to generate
corresponding GATEs.  On most drum synthesizers, you can plug plain old
audio into the TRIG IN jack, and it will trigger, so in this way it is
also a kind of "poor man's envelope follower". A passive 1U panel with
four drum triggers on it -- 1" piezo disks sandwiched between squares of
old mouse pad and quick-dry epoxy -- and four jacks, would be nice. That
way you can play your MOTM by thumping on it.

GATE time under voltage control? Nah, I think that should be a feature
of the envelope generator.

Starting point: http://www.paia.com/drumsens.htm. Honestly if I were
half as good with electronics as I am with software I'd try to build
this circuit, but I'm even stumped immediately by the power connector.
Seriously. Someone should take this soldering iron away from me before I
start to think that I know what I'm doing.

>I did a rough block diagram of this.  The implementation is 
>relatively easy up until you get to the part about eliminating the 
>crosstalk.  Believe me, I know what you guys are talking about.  I 
>have a set of V-Drums.
>
>The crosstalk issue could be dealt with one of two ways.  You could 
>compare the peak level of all four triggers and only allow the 
>largest, or you could compare the timing of the triggers and only 
>allow the first one.  I'm not sure if the propagation delay between 
>the pad being triggered for real and the other pads would be long 
>enough to make this practical.  In fact this is the feature most 
>likely to introduce latency.
>
>It's probably worth playing around with this.
>
>I'm thinking you have a very fast envelope follower.  As soon as the 
>incoming signal reaches a certain level a trigger is generated, 
>probably with a schmitt trigger.  The follower also drives a peak 
>detector.  The peak level is the CV.  A pot on the input controls 
>the sensitivity.  A pot on the output of the peak detector controls 
>the curve.  The trigger drives a gate with a variable on time. 
>Thats it for one pad.  You do have to decide when to drop the peak 
>in preparation for the next peak.  Possibly the leading edge of the 
>incoming trigger could reset the peak detector.
>
>For four pads you have to either compare each peak to the other 
>three and only the largest would generate a gate and trigger.  Or 
>you have the first trigger recieved lock out the other three for a 
>set time interval.  It would take some experimentation to determine 
>the best method.
>
>Food for thought.  How about it Scott.  Any way to MUUB this one?

Taking a trigger and converting it to a gate can be done with the 
TLN-866 gate processor. But I've never tried it with any kind of 
V-drum or piezo trigger, so I can't say if the levels from the drum 
pad are hot enough. Of course, you only need to change a couple of 
resistors to make it work with smaller signal levels.

It is possible to make a peak detector and crude envelope follower 
using MUUBs. But I think if you wanted a really good circuit you'd 
have to design a really good circuit and not limit yourself to what 
can be built using MUUBs, particularly when you want to minimize 
crosstalk. It's definately worth trying.

Wasn't Paul S. going to have an MOTM quality envelope follower this 
year? Could it also be used with this type of drum pad?

I would suspect that for best performance, the drum-trigger interface would be 
tailored for a specific type of pads.

Paul S.

On Mon, 2006-01-16 at 21:15 -0600, Paul Schreiber wrote:
> I would suspect that for best performance, the drum-trigger interface would be 
> tailored for a specific type of pads.

Most pads are just big things that let you hit them hard and thus
squeeze a piezo without breaking it.  If you have pads made with FSRs,
they invariably come with a MIDI host of some kind.

There are "mesh" head electronic drums and rubbery ones, but they both
seem to output the same thing.  The mesh ones are just more bouncy and
easier to use real snare technique on.

I'd be happy to lend/ship a few pintech pads and rubber pads to anyone
patient enough to work out this module and hold my hand in building a
few.  I'd even pitch in for materials if it helped.

-- 
(jfm3)

On Mon, 2006-01-16 at 21:07 -0500, Greg Amann wrote:
> [...] my sexy Roland meshy pads (they are stereo and use a bizarre
> technology that I have not spent the time to figure oot)

I'm not sure, but it's probably just two piezos that they analyze
cleverly and simultaneously in the host.  What happens if you plug them
into a trusty D4?

-- 
(jfm3)

Late night, I-should-already-be-asleep thoughts on reducing crosstalk:

Assumptions: For any given drum, there will be 1 or 2 other drums which
cause the majority of cross-talk. If this is completely wrong, never mind!
But...

If my assumption is reasonably correct, then each drum processor need 3
inputs: One normal input and 2 out-of-phase inputs with pots or trimpots.
Each drum processor also needs at least 3 outputs, in order to feed a normal
input and some anti-crosstalk inputs.

The anti-crosstalk stuff could all be behind the front panel, I guess, but
I'd rather see it up front. Large front panel, I know.

So, would that work?

(My first thoughts were of a bandpass filter, or a LP/HP pair of filters,
which would certianly be useful, anyway. Then I thought of the above.)
--
john

On Tue, 2006-01-17 at 00:32 -0500, john mahoney wrote:
> Late night, I-should-already-be-asleep thoughts on reducing crosstalk:
> 
> Assumptions: For any given drum, there will be 1 or 2 other drums which
> cause the majority of cross-talk. If this is completely wrong, never mind!
> But...

This is basically true, and a useful assumption.

> If my assumption is reasonably correct, then each drum processor need 3
> inputs: One normal input and 2 out-of-phase inputs with pots or trimpots.
> Each drum processor also needs at least 3 outputs, in order to feed a normal
> input and some anti-crosstalk inputs.
> 
> The anti-crosstalk stuff could all be behind the front panel, I guess, but
> I'd rather see it up front. Large front panel, I know.
> 
> So, would that work?

Yeah, I think it would work. It would probably be sufficient to have
crosstalk connections for each of your two neighbors on the bar you're
bolting the drums to.  

I feel it would be very important to keep the front panel 1U though,
since even a small kit must have several of these in order to be useful
(as opposed to, for example, a preamp module that you plug your guitar
into, of which you'd only need one).  If you wanted to build a modular
that had enough to reproduce some drum synth, you're looking at a
minimum of four, usually five or six "voices."  Each voice would need
one of these piezo->trigger/gate/cv things, an EG, and a filter...
that's already two panels for five voices, and we haven't even
considered noise sources, ring modulators, or stuff for more usual synth
voices. Besides, the only application of the crosstalk negating
connectors/trims would be for eliminating crosstalk.  It seems like
you'd have to get way too creative to come up with a more general use
for them, such that you'd want to easily connect them to other modules.

-- 
(jfm3)

Argh, the "Tellun pot format"!! Babel, here we come.

(Sorry Scott J, peace Greg, free country, DIY, each his own, whatever,
MOTM and Frac sleeping together, just kidding -- NOT!)

skw

On Jan 17, 2006, at 10:42 AM, Greg Amann wrote:

> Here's the deal: electronic drums (like Roland V-Drums and Yamaha DTX 
> etc.) are primarily performance instruments.  Modular synthesizers, 
> traditionally, are studio instruments.  So in trying to make a trigger 
> pad input module thingie, we really need to define the goal first.  
> For me, since I use my modular as a performance instrument, I am only 
> interested in ways to use my Roland V-Drum pads to make really cool 
> live music.
>
> It occurs to me that the following approaches might be considered:  
>
> 1.- Reverse engineer the trigger pad input circuitry on a Roland TD-20 
> into the modular format.  Paul H has certain reservations aboot some 
> of Roland's stuff, so reverse engineer an Alesis DM5 instead.  Or 
> better yet, just offer a conversion kit: I buy a DM5 off of eBay, buy 
> the Stooge conversion kit and away I go!  Woo hoo!
>
> 2.- Take a completely different approach.   Consider that burying the 
> crosstalk trimmers inside a case makes them set'n'forget: this is 
> completely useless to a performance instrument.  Every stage set-up 
> will require some tweaking.  And given the costs involved in the 
> inevitable module requirements to make this thing fun, it is just as 
> cost effective and more flexible to buy separate stands for each 
> pad (I often use cymbal stands with Gibraltar ball joint tom heads in 
> place of the cymbal arm) to achieve a brute force isolation than to 
> chew up 2U of panel space for each input module with the crosstalk 
> stuff on the front.
>
> Just my thoughts.
>
> I personally would way rather see a 1U with the Tellun small pot 
> format and the crosstalk accessible.
>
> PLL, BFG
>
>
>
>
> On 17-Jan-06, at 10:37 AM, jfm3 wrote:
>
>>  On Tue, 2006-01-17 at 00:32 -0500, john mahoney wrote:
>>  > Late night, I-should-already-be-asleep thoughts on reducing 
>> crosstalk:
>>  >
>>  > Assumptions: For any given drum, there will be 1 or 2 other drums 
>> which
>>  > cause the majority of cross-talk. If this is completely wrong, 
>> never mind!
>>  > But...
>>
>>  This is basically true, and a useful assumption.
>>
>>  > If my assumption is reasonably correct, then each drum processor 
>> need 3
>>  > inputs: One normal input and 2 out-of-phase inputs with pots or 
>> trimpots.
>>  > Each drum processor also needs at least 3 outputs, in order to 
>> feed a normal
>>  > input and some anti-crosstalk inputs.
>>  >
>>  > The anti-crosstalk stuff could all be behind the front panel, I 
>> guess, but
>>  > I'd rather see it up front. Large front panel, I know.
>>  >
>>  > So, would that work?
>>
>>  Yeah, I think it would work. It would probably be sufficient to have
>>  crosstalk connections for each of your two neighbors on the bar 
>> you're
>>  bolting the drums to.
>>
>>  I feel it would be very important to keep the front panel 1U though,
>>  since even a small kit must have several of these in order to be 
>> useful
>>  (as opposed to, for example, a preamp module that you plug your 
>> guitar
>>  into, of which you'd only need one). If you wanted to build a modular
>>  that had enough to reproduce some drum synth, you're looking at a
>>  minimum of four, usually five or six "voices." Each voice would need
>>  one of these piezo->trigger/gate/cv things, an EG, and a filter...
>>  that's already two panels for five voices, and we haven't even
>>  considered noise sources, ring modulators, or stuff for more usual 
>> synth
>>  voices. Besides, the only application of the crosstalk negating
>>  connectors/trims would be for eliminating crosstalk. It seems like
>>  you'd have to get way too creative to come up with a more general use
>>  for them, such that you'd want to easily connect them to other 
>> modules.
>>
>>  --
>>  (jfm3)
>>
>>
>> YAHOO! GROUPS LINKS
>>
>> 	▪ 	 Visit your group "motm" on the web.
>>  
>> 	▪ 	 To unsubscribe from this group, send an email to:
>>  motm-unsubscribe@yahoogroups.com
>>  
>> 	▪ 	 Your use of Yahoo! Groups is subject to the Yahoo! Terms of 
>> Service.
>>
>>
>
>
>
>
> SPONSORED LINKS
> Label
> Cd music label
> Independent music label
>
> YAHOO! GROUPS LINKS
>
> 	▪ 	 Visit your group "motm" on the web.
>  
> 	▪ 	 To unsubscribe from this group, send an email to:
>  motm-unsubscribe@yahoogroups.com
>  
> 	▪ 	 Your use of Yahoo! Groups is subject to the Yahoo! Terms of 
> Service.
>
>

>Argh, the "Tellun pot format"!! Babel, here we come.
>

Careful, I don't want the mounties knocking on my door. ;-)

--- In motm@yahoogroups.com, Greg Amann <greg.amann@s...> wrote:
> 
> I personally would way rather see a 1U with the Tellun small pot  
> format and the crosstalk accessible.

Another possibility might be to use a pot with a slotted shaft (or a
high quality trim pot) and have a couple of small round holes on the
panel for the crosstalk trims.  (In other words, a small screwdriver
would be required to adjust the trims but you could do it from the
front panel.)  This should allow for a 1U panel with big knobs for the
more commonly tweaked controls.  The crosstalk inputs and outputs
would be jacks on the back side of the module and as such would be
semi-permanently patched.

This actually sounds like a pretty fun module.  I think I need some of
these too.

seth

This is further justification for building whatever drum trigger
interface to support the model of "a piezo disk wrapped in something you
whack."

I did some reading on http://www.harmony-central.com/Drums/EDW/.  It
seems like your fancy pads will either have two piezos or one piezo and
a switch. I have one of these switched pads myself -- the idea is you
hit the pad and then grab it with your other hand as if you were choking
a cymbal.  I haven't seen a lot of electronic drummers who demanded to
be able to do cymbal chokes this way, they're just as happy to hit a
different pad or pedal or whatever.  Either way, if you have a module as
proposed so far, you're set.  You just might need two piezo->trig/gate
things per head, and perhaps a little adapter that goes from one TRS
phone plug to two TS phone jacks (not hard to build or find at Radio
Shack) so you can get both of the piezo signals or the piezo signal and
the switch into plain old TS 1/4" phone plug terminated cable.

It's all piezo disks in any manufacturer's electronic drums, except for
some of the integrated controllers that use FSRs like the DrumKat.
DrumKats are still in production, but relatively expensive, and there's
no hope of hacking one to put out anything but MIDI.  The FSRs used in
the Alternate Mode products can last for as little as 6 months of
regular playing, production quality varies, and they cost hundreds to
replace, replacements only being available from the manufacturer.
Piezos last somewhat longer, can be replaced for less than US$2, and are
available at Radio Shack.  FSRs generate a much cleaner pulse output
when you whack them, extremely few crosstalk and false triggering
problems, and you can also get "aftertouch" if you push your stick into
them.  It's amazing to see Mario play them.  www.alternatemode.com

You can buy a bunch of Remo practice pads, epoxy some piezos to coffee
can lids and shove them inside, and so have drum triggers as good as any
*and* delicious home brewed coffee.  Or, you can go buy mesh heads with
exotic foamy cones to transfer energy from the head to the piezo, and
drum shells made of some kind of "acoustically dead polymer".  Either
way, they give out basically the same signal.

On Wed, 2006-01-18 at 07:54 -0500, Greg Amann wrote:
> Well they have stereo 1/4" plugs so you get the centre pad but the  
> rim is shorted oot.
> 
> PLL, BFG
> 
> On 16-Jan-06, at 11:51 PM, jfm3 wrote:
> 
> > On Mon, 2006-01-16 at 21:07 -0500, Greg Amann wrote:
> >> [...] my sexy Roland meshy pads (they are stereo and use a bizarre
> >> technology that I have not spent the time to figure oot)
> >
> > I'm not sure, but it's probably just two piezos that they analyze
> > cleverly and simultaneously in the host.  What happens if you plug  
> > them
> > into a trusty D4?
-- 
(jfm3)

hello all,
since we're talking piezo based drumpads - this is my little DIY page  
on these things
http://www.unrecordings.co.uk/drumpad.html
cheers
paul b (sheffield uk)